The caller begins to ask a question concerning the terminal at which a particular flight will arrive. Before this turn is complete, the agent is able to predict that
the next TRP will come at the end of the word ‘terminal’. Evidence for this comes from the agent’s in-breath after the initial sound of the word ‘terminal’,
indicating she is gearing up to speak. She begins her next turn right at the anticipated end of the turn construction unit ‘which terminal’. However, the
caller’s addition of a further component ‘is it’ at the TRP results in a short period of overlapping talk.
2.10 From Wooffitt
et al, 1997: 115
127 A:
does he know you, 128
C: yes he does he does 129
A: right
Here, the overlap arises because the agent initiates her turn at the transition relevance place following the turn construction unit ‘yes he does’, at the same
time that the caller repeats her last two words.
2.11 From Wooffitt
et al, 1997: 109
10 A:
yes the four three one from amsterdam 11
came in at thirteen oh five madam 12
1 13
C: thirteen oh five l ovely thank you very
14 A:
that’s right 15
C: much indeed for your help
Here in line 13 the caller repeats the information the agent has provided. Pretty much at the anticipated end of the turn construction unit in which the repeat
is done, the agent confirms the information, and thus overlaps with the caller who has now initiated another turn construction unit to thank the agent.
Jefferson has conducted extensive analyses of the organisation of overlap- ping talk Jefferson, 1983, 1986. The clear finding from her studies is that
instances of overlap are either the result of next speakers starting in anticipa- tion of the forthcoming transition relevant place, or the consequence of speak-
ers orienting to the relevance of different aspects of the rule set identified by Sacks
et al. Routinely, then, instances of overlapping talk transpire to be an orderly consequence of that system, not a deviation from it.
The turn-by-turn development of interaction is not simply a series of utter- ances coming one after another from different participants. There are connections
between turns which yield describable and consistent properties. To illustrate this we need to go back to Rule 1a of the model outlined by Sacks
et al, and con- sider one powerful method by which a current speaker can select a next speaker.
TWO KEY STUDIES 31
关
关
In extract 2.5, S asks the question, ‘Oscar did you work for somebody before you worked for Zappa?’ and thereby issues the first part of a question–answer
paired action sequence. Intuitively, it seems that some kinds of conversational actions belong with
each other. Greetings, such as ‘hi’–‘hi’, seem to form a ‘natural’ pair. It also seems natural that questions will be followed by answers, and that offers will
be followed by acceptances or refusals, and so on. Right from the start of his studies of interaction Sacks was interested in these kinds of paired units. To
provide a formal account of their generic properties, Sacks proposed the con- cept of the
adjacency pair more formally outlined in Schegloff and Sacks, 1973. Heritage 1984a provides the following formulation. An adjacency
pair is a sequence of two utterances which are adjacent, produced by differ- ent speakers, ordered as a first part and second part and typed, so that a first
part requires a particular second, or range of second parts Heritage, 1984a: 246. An invitation, then, would be the first part of an invitation–response
pair, a question the first part of a question–answer pair, and a greeting the first part of a greeting–greeting pair.
There is a normative relationship between the turns that constitute paired sequences. A speaker’s production of the first part of a pair generates the
expectation that an allocated next speaker should produce the appropriate
second part. The second part of a pair is said to be conditionally relevant after the production of a first part Schegloff, 1972a. So, if a next speaker is
selected via the first part of a pair, not only are they obliged to speak, but they will be expected to provide the appropriate second pair part, or an account for
its absence.
Heritage 1984a: 247–53 discusses various kinds of evidence that partici- pants in interaction are sensitive to these expectations. We will discuss just
one: what happens when a first part of a pair has been produced, but the appropriate next part is not forthcoming. In extract 2.12, a child has asked her
mother a question. After a gap of over a second, the mother has not answered, and the child speaks again.
2.12 From Atkinson and Drew, 1979: 52; discussed in Heritage, 1984a:
248–9
Child: Have to cut the:se Mummy
1.3 Child:
Won’t we Mummy 1.5
Child: Won’t we
Mother: Yes
Had the child interpreted the mother’s silence as indicating that she hadn’t heard the question, it is likely that it would have been repeated, perhaps louder.
Instead, the child provides increasingly truncated versions of the initial question.
32 CONVERSATION ANALYSIS AND DISCOURSE ANALYSIS
